Sören Balko

Design Support for Database Federations

Federated database systems provide a homogeneous interface to possibly heterogeneous local database systems. This homogeneous interface consists of a global schema which is the result of a logical integration of the schemata of the corresponding local database systems and file systems. In this paper, we sketch the integration process and a set of tools for supporting the design process. Besides the classical database schema integration, the design process for federated information systems requires the integration of other aspects like integrity rules, authorization policies and transactional processes. This paper reports on an integrated approach to tool support of several of these integration aspects. The different integration facets are linked via the database integration method GIM allowing a high degree of automatic integration steps.

The Active Vertice method: a performant filtering approach to high-dimensional indexing

The problem of finding nearest neighbors has emerged as an important foundation of feature-based similarity search in multimedia databases. Most spatial index structures based on the R-tree have failed to efficiently support nearest neighbor search in arbitrarily distributed high-dimensional data sets. In contrast, the so-called filtering principle as represented by the popular VA-file has turned out to be a more promising approach. Query processing is based on a flat file of compact vector approximations. In a first stage, those approximations are sequentially scanned and filtered so that in a second stage the nearest neighbors can be determined from a relatively small fraction of the data set.In this paper, we propose the Active Vertice method as a novel filtering approach. As opposed to the VA-file, approximation regions are arranged in a quad-tree like structure. High-dimensional feature vectors are assigned to ellipsoidal approximation regions on different levels of the tree. A compact approximation of a vector corresponds to the path within the index from the root to the respective tree node. When compared to the VA-file, our method enhances the discriminatory power of the approximations while maintaining their compactness in terms of storage consumption. To demonstrate its effectiveness, we conduct extensive experiments with synthetic as well as real-life data and show the superiority of our method over existing filtering approaches.

BioDataServer: an Applied Molecular Biological Data Integration Service

Nowadays, huge volumes of molecular biological data are available from different biological research projects. This data often covers overlapping and complemental domains. For instance, the Swiss-Prot database merely contains protein sequences along with their annotations, whereas the KEGG database incorporates enzymes, metabolic pathways and genome data. Due to the fact that this data complements and completes each other, it is desirable to gain a global view on the integrated databases instead of browsing each single data source itself.

A service-oriented grid infrastructure for multimedia management and search

Nowadays, digital libraries are inherently dispersed over several peers of a steadily increasing network. Dedicated peers may provide specialized, computationally expensive services such as image similarity search. Usually, the peers of such a network are uncoordinated in the sense that their content and services are not linked together. Nevertheless, users expect to transparently access and modify all the (multimedia) content anytime from anywhere not only in an efficient and effective but also consistent way. To match these demands, future digital libraries require an infrastructure that combines various information technologies like databases, service-oriented architectures, peer-to-peer and grid computing. In this paper, we sketch such an infrastructure and illustrate how an example digital library application can work atop it. In detail, we show how similarity search is supported by this infrastructure, and discuss how query distribution and load balancing based on domain-specific knowledge can be exploited by the infrastructure to reduce query response times.

Towards Enhanced Compression Techniques for Efficient High-Dimensional Similarity Search in Multimedia Databases

In this paper, we introduce a new efficient compression technique for high-dimensional similarity search in MMDBS. We propose the Active Vertice Tree which is based on concave cluster geometries. Furthermore, we briefly sketch a model for high-dimensional point alignments and specify basic requirements for high-dimensional cluster shapes. Finally, we compare the Active Vertice Tree with other methods for high-dimensional similarity search in terms of their retrieval performance.

Compression Techniques Based on Concave Cluster Geometries for Efficient High-Dimensional Nearest Neighbor Retrieval

In this paper, we discuss the problem domain of high-dimensional nearest neighbor retrieval. We give a brief overview on existing approaches based on convex cluster shapes. Subsequently, we sketch the advantage of concave cluster geometries and introduce three concave cluster proposals. Furthermore, we put our concave clustering approaches into a context with index compression techniques. Finally, an outlook on our ongoing work concludes this paper.